On June 4, wireless chip giant Qualcomm (NASDAQ:QCOM) announced a new mobile platform called the Snapdragon 850. The Snapdragon 850, Qualcomm says, is aimed at so-called "always connected" personal computers -- that is, Windows-based notebook computers that have cellular data connections.
Qualcomm cites some impressive figures, claiming 30% better "systemwide" performance, a tripling in artificial intelligence performance, support for 1.2 gigabit LTE speeds, and "up to 25 hours of continuous usage or multi-day battery life under normal usage conditions" -- with, of course, the caveats that "battery life varies significantly with settings, usage, and other factors."
Nevertheless, while Qualcomm aggressively promoted its predecessor (the Snapdragon 835) and is now trying to make a splash with this part, I'm convinced that it's doomed to fail. Here are three reasons why.
1. It's still ARM-based
The overwhelming majority of Windows-based computers sold today incorporate processors that are compatible with Intel (NASDAQ:INTC) architecture, commonly referred to as x86. For decades, the Windows software ecosystem has grown up around Intel architecture-based processors.
The Snapdragon 850 is based on the ARM architecture. While the ARM architecture dominates the smartphone market, it hasn't found much success in the personal computer or data center markets (both areas that ARM and its partners continue to attack).
The problem with the Snapdragon 850 being an ARM-based processor is simple: It won't run any software that hasn't been specifically rewritten for the ARM architecture at anywhere near the speeds that even a low-end Intel Celeron or Pentium processor will because of the software emulation overhead.
This is just a disadvantage that any ARM-based processor running on Windows is going to have to deal with for the foreseeable future.
This issue is exacerbated by the fact that Qualcomm's Snapdragon Mobile platforms for Windows are based on the same basic chip technology as the processors that it sells into premium smartphones. These processors continue to get faster and more efficient with time, but the level of performance that something like the Snapdragon 850 is going to offer is going to be much lower than a competing Intel-based part designed specifically for notebooks.
2. Weak ecosystem
Although Qualcomm has partnered with major PC makers like ASUS, HP, and Lenovo to put out notebooks based on its Snapdragon processors, the reality is that there are very few computer designs from these vendors based on Qualcomm's chips.
In fact, I can only find one model based on the Snapdragon 835 from each of the aforementioned laptop makers.
By contrast, whenever Intel releases a new processor platform, there are usually north of 100 systems based on that platform that come out. For example, Intel recently announced its Amber Lake-Y and Whiskey Lake-U mobile processors and said that we should "expect more than 140 new laptops and 2 in 1s from OEMs starting this fall."
To make matters worse, Qualcomm has no presence in the desktop processor market, so its potential volume scale with system makers is naturally hamstrung.
I'm not done yet, though. Intel invests billions of dollars each year in marketing -- the company's marketing, sales, and administrative expenses was $7.47 billion in 2017. Much of that marketing spending is related to the company's personal computer processor business.
Qualcomm's total sales, general and administrative expenses during its fiscal 2017, by contrast, was just $2.66 billion. Most of that spending is likely applied to its wireless chip business, which generated nearly $16.5 billion in sales in that time.
If Qualcomm wants to seriously compete with Intel, not only does it need to overcome the inherent technology disadvantages of its products in Windows-based computers, but it must fight Intel's huge marketing muscle, too.
Is Qualcomm really ready to make that investment? I doubt it.
3. Narrow product portfolio
Notice that Qualcomm announced a single product for the Windows notebook market -- the Snapdragon 850. Contrast this with Intel, which has a very broad portfolio of targeted solutions for each subsegment of the personal computer market. Not only does Intel have different products for each subsegment, but it has multiple models within each processor family to allow computer makers to offer lower-cost and higher-end models to address a wide range of price points.
For example, starting this fall, Intel's personal computer processor portfolio will include the following:
- Gemini Lake for low-cost notebooks and all-in-one desktops.
- Coffee Lake-S for the all-in-one and traditional desktop markets.
- Coffee Lake-H for high-performance notebooks.
- Coffee Lake-U for high-end mainstream notebooks.
- Whiskey Lake-U for mainstream notebooks.
- Amber Lake-Y for tablet/laptop hybrids and very low-power, high-end notebooks.
Qualcomm will have, presumably, the Snapdragon 850 platform. That's it.
In terms of product portfolio breadth and depth, Qualcomm's offerings can't come close to Intel's, so Qualcomm's total addressable market is extremely limited and I suspect that it won't capture a large percentage of that total addressable market.
The Snapdragon 850, like its predecessor, is set to be a commercial failure. Qualcomm needs to stop wasting shareholder money trying to chase this market.